Developed an intelligent system to convert handwritten English text into machine-readable format using a CRNN architecture (ResNet18 + BiLSTM + CTC Loss). The model handles diverse handwriting styles without character segmentation and achieves high accuracy. Deployed as a Flask-based web application for real-time predictions. #DeepLearning #MachineLearning #ComputerVision #OCR #Python #AI #StudentProject

ODSC AI East 2026 is coming up, and Hydrolix Principal TAM Dan Sullivan will be speaking on a topic that will resonate with teams building decision support for complex systems. In “Spec-Driven Simulation Modeling: Building and Validating Decision Support Models with Python and LLMs,” Dan explores why simulation modeling can be the better fit when prediction alone does not capture queues, constraints, and feedback loops. Register here: https://hubs.la/Q049YCDX0 #Hydrolix #ODSCAIEast #DataScience #MachineLearning #MLOps #SimulationModeling #Python #LLMs

I recently developed a #computational workflow bridging #Python and #Rhino to operationalize the Geometrical Potential for Daylighting (#GPD) method. Through a custom Python script embedded in Rhino, #GPD can now be calculated in just a few intuitive steps, transforming what was once a conceptual framework into a fast, repeatable #design tool. In parallel, I introduced #sGPD as a spatial extension of the method, generating real-time heatmaps across floor plates to visually communicate performance variations. This integration not only simplifies analysis but turns #geometry into an active driver of insight, enabling designers to evaluate and iterate spatial performance with clarity, speed, and precision.

Object Detection & Tracking (Computer Vision) Best paired with: The final_tracked_output.mp4 video showing the bounding boxes. Bringing real-time AI to life! 👁️💻 I recently completed a new Computer Vision pipeline focused on Object Detection and Tracking. Using YOLOv8, I set up a system that can process video feeds, instantly detect objects with high confidence, and—most importantly—track them. By integrating deep tracking algorithms (SORT/ByteTrack), the model doesn't just see a "person"; it assigns a unique tracking ID to that specific individual and follows them seamlessly across frames. Building this in Python and handling the complexities of video processing and frame-by-frame tracking was an incredible learning experience. Check out the model in action below! 👇 CodeAlpha #ComputerVision #YOLOv8 #MachineLearning #ArtificialIntelligence #Python #ObjectDetection #TechJourney

Claude now connects to Blender With the new Blender connector, Claude can work directly inside your 3D workflow from debugging scenes to building custom tools and applying changes across multiple objects. By using natural language with Blender’s Python API, it turns complex operations into simple instructions, helping automate repetitive tasks and speed up production. A step toward AI becoming a true collaborator inside creative tools not just outside them. #AI #Claude #Blender #3D #Automation

This design transforms the Python ecosystem into a cinematic, futuristic infographic. Instead of a flat pyramid, it features a cybernetic “data snake” winding its way through a digital universe. Each section is represented by a high-impact visual icon: Machine Learning: A glowing brain inside a golden polyhedron. Web Development: A connected global node representing the web. Automation Testing: Precision mechanical gears. Image Processing: A high-tech camera lens. Game Development: A game controller with a retro-futuristic aesthetic. Web Scraping: A magnifying glass analyzing a network of information. The use of blue, gold, and red neon lights against a deep-space background creates a clear visual hierarchy, transforming a technical list into a piece of digital art. #Python #Programming #MachineLearning #WebDev #CodingLife

Upgraded the backend architecture for my Naruto Hand-Sign Recognition project. Here's what I changed: - Unified feature extraction across live runtime + prediction API - Added shared signals: landmark distances, angles, motion features - Introduced per-class confidence thresholds for cleaner outputs - Improved prediction stability with stronger decision logic - Preserved custom rule-based Rasengan detection path - Cleaner training pipeline + better debugging visibility Result: more accurate predictions, fewer false triggers, and a stronger base for future jutsu detection. #ComputerVision #MachineLearning #OpenCV #MediaPipe #Python #AI #BackendDevelopment #Naruto #BuildInPublic

New tutorial | Run Inference on web browser using Ultralytics YOLO26 + Streamlit 🌐 Building web-based AI applications is easier than ever with tools like Streamlit, especially when combined with real-time computer vision models. In this tutorial, we show how to create a live inference web app using #YOLO26 and #Streamlit, enabling real-time vision directly in the browser. What'll be covered: ✅ Streamlit live inference solution overview ✅ Running the application in Python ✅ Object detection in a web interface ✅ Instance segmentation in real time ✅ Processing video inputs for live predictions A practical guide to building interactive, browser-based Vision AI applications with YOLO26. Watch now ➡️ https://bit.ly/4sYIqNp #Ultralytics #YOLO26

Excited to share a new project from my portfolio: Clue2 — a Python implementation of the classic deduction game Clue, powered by a logic‑driven AI detective. What I built: A modular game engine and state tracker for full game flow A propositional‑logic knowledge base that lets the AI reason about cards and opponent reveals Simulation tools to run automated games and measure agent performance A full suite of unit tests to keep the system reliable and maintainable What I learned: Practical propositional reasoning & knowledge representation Designing deterministic simulations for agent evaluation Writing reliable, test‑first Python for game systems If you’re curious about the design or want a walkthrough, feel free to DM me or check out the repo. Always open to feedback and opportunities involving AI, simulation, or systems engineering. GitHub link: https://lnkd.in/ggyTarrU #Python #AI #GameDev #Logic #TDD #OpenSource #Portfolio

Automation Case 2: Well Pattern Allocation Another practical automation case from my reservoir-engineering work. This one is about a routine but important step behind pattern-based VRR analysis: assigning which injectors support which producers, and with what share. I turned this into a simple Python workflow that: 🔹reads wells’ coordinates and well types, 🔹calculates bottom-hole distances automatically, 🔹assigns injector support shares by geometry rules, 🔹exports a ready allocation table and schematic map. Result: a task that usually takes hours of manual mapping becomes faster, more consistent, and repeatable. #ReservoirEngineering #Waterflood #Python #AI